Tea samples from the Polish market were analysed for the determination of pyrrolizidine alkaloids (PAs). The method applied in the study is based on the GC-MS technique and uses the sum parameter approach for quantification. The analytical protocol was subjected to validation in terms of linearity, repeatability, reproducibility, the limit of quantification and determination, matrix effect, selectivity, and uncertainty, and all evaluated parameters fulfilled the requirements of Commission Decision 2002/657. A total of 78% of the analysed samples contained 1,2-unsaturated PAs, and the concentrations were determined to fall in the range of 2–6498 µg kg−1. The highly contaminated samples were additionally subjected to LC-MS analysis. Much higher PA concentrations were determined with the sum parameter method, proving its superior quantification capability. The high contamination found prompted an assessment of the dietary exposure from the consumption of infusions prepared from the analysed teas. Two approaches were adopted: one assuming 100% transfer of PAs to the infusions and the second assuming only partial transfer. Regardless of the adopted approach, consumption of several infusions was shown to greatly exceed the maximum acceptable daily intake of PAs.